varY <- "Bardella"
nameY<-paste("% vote", varY)Autocorrelation spatiale
Application aux résultats des élections européennes de 2024 en France
- Mise en place : Télécharger le dossier exo_autocor.zip et décompressez le sur votre ordinateur. Puis créez un projet R avant d’executer le programme
Vous allez choisir une des listes électorales puis reprendre l’ensemble des analyses vues en cours à l’aide des programmes ci-dessous. Les choix possibles sont :
“Aubry” “Bardella” “Bellamy” “Deffontaines” “Glucksman” “Hayer” “Lassale” “Marechal” “Toussaint” “Autres”
A. ECHELLE REGIONALE
Préparation des données
### Fonds de carte
map<-st_read("data/elect2024/map_reg.shp", quiet=T) %>%
select(code=code_reg, geometry) %>%
st_transform(2154)
### Données
don<-readRDS("data/elect2024/don_reg.RDS") %>%
select(code=code_reg, nom=nom_reg, 3:12)
don$nom<-as.factor(don$nom)
levels(don$nom)<-c("ACAL", "AQUI","AURA","BOFC","BRET","CVDL","IDF","OCCI","NOPI","NORM","PDL","PACA")
### Jointure
mapdon<-left_join(map[,1],don)
# Table de voisinage
map_nb<-spdep::poly2nb(mapdon,row.names = mapdon$code)
# Table de poids
map_nb_w<-nb2listw(map_nb)
summary(map_nb_w)Characteristics of weights list object:
Neighbour list object:
Number of regions: 12
Number of nonzero links: 46
Percentage nonzero weights: 31.94444
Average number of links: 3.833333
Link number distribution:
2 3 4 5 6
2 3 3 3 1
2 least connected regions:
53 93 with 2 links
1 most connected region:
24 with 6 links
Weights style: W
Weights constants summary:
n nn S0 S1 S2
W 12 144 12 6.643889 49.32222# Carte de voisinage
coo<-st_coordinates(st_centroid(mapdon))
mf_map(mapdon, type="base",col="lightyellow")
mf_layout("Carte des liens de voisinage", frame=T)
plot.nb(map_nb,coords = coo,add = T,col = "red",points = F)
mf_label(mapdon, var="code", col="blue",cex=0.6,halo = T, bg="white",r = 0.1)
Choix d’une variable Y
sel<-mapdon %>% select(code,nom, Y=varY) %>%
mutate(Y_std = as.numeric(scale(Y)) ) par(mfrow=c(1,2))
# Carte de Y
mf_map(sel, var="Y", type="choro",
nbreaks = 6,
leg_title = "Quantiles",
leg_val_rnd = 1)
mf_layout("Y",
frame=T,
credits="",
arrow=F)
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Y_std",
frame=T,
credits="",
arrow=F)
Valeurs de voisinage
sel$Y_lag<-as.numeric(lag.listw(map_nb_w,sel$Y))
sel$Y_std_lag<-as.numeric(lag.listw(map_nb_w,sel$Y_std))
#mapsel$Y_std_lag<-lag.listw(map_nb_w,mapsel$Y_std)
st_drop_geometry(sel) code nom Y Y_std Y_lag Y_std_lag
1 11 IDF 18.78888 -2.1535547 37.61953 0.72909500
2 24 CVDL 34.93906 0.3187601 30.11507 -0.41970883
3 27 BOFC 37.08970 0.6479876 30.74364 -0.32348574
4 28 NORM 35.32945 0.3785223 29.86869 -0.45742619
5 32 NOPI 42.40672 1.4619321 30.81702 -0.31225318
6 44 ACAL 38.33273 0.8382729 32.76177 -0.01454500
7 52 PDL 27.62945 -0.8002157 31.69672 -0.17758538
8 53 BRET 25.57934 -1.1140528 31.47945 -0.21084671
9 75 AQUI 30.93906 -0.2935711 31.79258 -0.16291148
10 76 OCCI 33.68791 0.1272309 33.49939 0.09837241
11 84 AURA 30.91391 -0.2974213 35.06019 0.33730344
12 93 PACA 38.64522 0.8861097 32.30091 -0.08509518Indice et diagramme de Moran
spdep::moran.test(sel$Y,map_nb_w, alternative = "two.sided")
Moran I test under randomisation
data: sel$Y
weights: map_nb_w
Moran I statistic standard deviate = -0.69097, p-value = 0.4896
alternative hypothesis: two.sided
sample estimates:
Moran I statistic Expectation Variance
-0.20852369 -0.09090909 0.02897351 par(mfrow=c(1,1))
moran.plot(x=sel$Y_std,
listw=map_nb_w,
pch=20,
main ="Diagramme de Moran",
xlab = "Variable observée (Y)",
ylab= "Variable de voisinage (Ylag)",labels = F)
text(sel$Y_std,
sel$Y_std_lag,
sel$nom,
cex=0.5,
col="red",
pos=1)
Calcul des LISA
locm<-localmoran(sel$Y,map_nb_w,alternative = "two.sided")
sel2<-as.data.frame(locm)
tabres<-cbind(sel,sel2)
st_drop_geometry(tabres) code nom Y Y_std Y_lag Y_std_lag Ii E.Ii
1 11 IDF 18.78888 -2.1535547 37.61953 0.72909500 -1.71288650 -0.459946894
2 24 CVDL 34.93906 0.3187601 30.11507 -0.41970883 -0.14594885 -0.010076829
3 27 BOFC 37.08970 0.6479876 30.74364 -0.32348574 -0.22867063 -0.041641779
4 28 NORM 35.32945 0.3785223 29.86869 -0.45742619 -0.18888653 -0.014209497
5 32 NOPI 42.40672 1.4619321 30.81702 -0.31225318 -0.49799230 -0.211958224
6 44 ACAL 38.33273 0.8382729 32.76177 -0.01454500 -0.01330111 -0.069689401
7 52 PDL 27.62945 -0.8002157 31.69672 -0.17758538 0.15502539 -0.063505302
8 53 BRET 25.57934 -1.1140528 31.47945 -0.21084671 0.25624840 -0.123085644
9 75 AQUI 30.93906 -0.2935711 31.79258 -0.16291148 0.05217394 -0.008547175
10 76 OCCI 33.68791 0.1272309 33.49939 0.09837241 0.01365383 -0.001605393
11 84 AURA 30.91391 -0.2974213 35.06019 0.33730344 -0.10944134 -0.008772835
12 93 PACA 38.64522 0.8861097 32.30091 -0.08509518 -0.08225854 -0.077870117
Var.Ii Z.Ii Pr.z....E.Ii..
1 0.357689878 -2.09496520 0.03617409
2 0.009975286 -1.36040224 0.17370267
3 0.083806257 -0.64605635 0.51824287
4 0.020170926 -1.22990878 0.21873126
5 0.534502193 -0.39123980 0.69561999
6 0.207464924 0.12379885 0.90147454
7 0.124891996 0.61836534 0.53633454
8 0.582852069 0.49686985 0.61928083
9 0.017795654 0.45517954 0.64898008
10 0.005129009 0.21306678 0.83127487
11 0.012522057 -0.89961308 0.36832619
12 0.387754354 -0.00704742 0.99437702Carte des LISA
par(mfrow=c(1,2))
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Variable standardisée",
frame=T,
credits="",
arrow=F)
mypal<-brewer.pal(5,"BrBG")
mybreaks<-c(-10,-2,-1,1,2,10)
mf_map(tabres, var="Z.Ii", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "Z-score",
leg_val_rnd = 1
)
mf_layout("LISA",
frame=T,
credits="",
arrow=F)
Typologie de Moran
# create a new variable identifying the moran plot quadrant for each observation, dismissing the non-significant ones
tabres$typ <- NA
# high-high quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-high"
# low-low quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-low"
# high-low quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-low"
# low-high quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-high"
# non-significant observations
tabres[tabres$Pr.z....E.Ii.. > 0.05, "typ"] <- "not signif."
tabres$typ <- as.factor(tabres$typ)
tabres$typcol<-NA
tabres$typcol <- case_when(tabres$typ=="high-high" ~ "red",
tabres$typ=="low-low" ~ "blue",
tabres$typ=="high-low" ~ "pink",
tabres$typ=="low-high" ~ "lightblue",
tabres$typ=="not signif." ~ "gray90"
)
mf_map(tabres, type="base", col=tabres$typcol)
mf_layout("Typologie de Moran",
frame=T,
credits ="")
mf_legend(type="typo",val = c("haut-haut","bas-bas","haut-bas","bas-haut","non sign."),pal = c("red","blue","pink","lightblue","gray90"),title = "Types")
B. ECHELLE DEPARTEMENTALE
Préparation des données
### Fonds de carte
map<-st_read("data/elect2024/map_dept.shp", quiet=T) %>%
select(code=code_dpt, geometry) %>%
st_transform(2154)
### Données
don<-readRDS("data/elect2024/don_dept.RDS") %>%
select(code=code_dpt, nom=nom_dpt, 5:14)
### Jointure
mapdon<-left_join(map[,1],don)
# Table de voisinage
map_nb<-spdep::poly2nb(mapdon,row.names = mapdon$code)
# Table de poids
map_nb_w<-nb2listw(map_nb)
summary(map_nb_w)Characteristics of weights list object:
Neighbour list object:
Number of regions: 94
Number of nonzero links: 474
Percentage nonzero weights: 5.364418
Average number of links: 5.042553
Link number distribution:
2 3 4 5 6 7 8 10
6 12 14 19 30 11 1 1
6 least connected regions:
06 29 57 62 66 74 with 2 links
1 most connected region:
77 with 10 links
Weights style: W
Weights constants summary:
n nn S0 S1 S2
W 94 8836 94 40.29206 383.9874# Carte de voisinage
coo<-st_coordinates(st_centroid(mapdon))
mf_map(mapdon, type="base",col="lightyellow")
mf_layout("Carte des liens de voisinage", frame=T)
plot.nb(map_nb,coords = coo,add = T,col = "red",points = F)
mf_label(mapdon, var="code", col="blue",cex=0.6,halo = T, bg="white",r = 0.1)
Choix d’une variable Y
sel<-mapdon %>% select(code,nom, Y=varY) %>%
mutate(Y_std = as.numeric(scale(Y)) ) par(mfrow=c(1,2))
# Carte de Y
mf_map(sel, var="Y", type="choro",
nbreaks = 6,
leg_title = "Quantiles",
leg_val_rnd = 1)
mf_layout("Y",
frame=T,
credits="",
arrow=F)
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Y_std",
frame=T,
credits="",
arrow=F)
Valeurs de voisinage
sel$Y_lag<-as.numeric(lag.listw(map_nb_w,sel$Y))
sel$Y_std_lag<-as.numeric(lag.listw(map_nb_w,sel$Y_std))
#mapsel$Y_std_lag<-lag.listw(map_nb_w,mapsel$Y_std)
head(st_drop_geometry(sel)) code nom Y Y_std Y_lag Y_std_lag
1 01 AIN 36.53036 0.3548758 31.39537 -0.3338932
2 02 AISNE 50.63966 2.2473937 40.51463 0.8892969
3 03 ALLIER 36.69576 0.3770614 35.98661 0.2819411
4 04 ALPES-DE-HAUTE-PROVENCE 36.30369 0.3244723 37.09966 0.4312381
5 05 HAUTES-ALPES 31.05594 -0.3794221 33.09124 -0.1064220
6 06 ALPES-MARITIMES 37.73249 0.5161209 39.76664 0.7889666Indice et diagramme de Moran
spdep::moran.test(sel$Y,map_nb_w, alternative = "two.sided")
Moran I test under randomisation
data: sel$Y
weights: map_nb_w
Moran I statistic standard deviate = 8.8275, p-value < 2.2e-16
alternative hypothesis: two.sided
sample estimates:
Moran I statistic Expectation Variance
0.56608333 -0.01075269 0.00427000 par(mfrow=c(1,1))
moran.plot(x=sel$Y_std,
listw=map_nb_w,
pch=20,
main ="Diagramme de Moran",
xlab = "Variable observée (Y)",
ylab= "Variable de voisinage (Ylag)",labels = F)
text(sel$Y_std,
sel$Y_std_lag,
sel$nom,
cex=0.5,
col="red",
pos=1)
Calcul des LISA
locm<-localmoran(sel$Y,map_nb_w,alternative = "two.sided")
sel2<-as.data.frame(locm)
tabres<-cbind(sel,sel2)
head(st_drop_geometry(tabres)) code nom Y Y_std Y_lag Y_std_lag
1 01 AIN 36.53036 0.3548758 31.39537 -0.3338932
2 02 AISNE 50.63966 2.2473937 40.51463 0.8892969
3 03 ALLIER 36.69576 0.3770614 35.98661 0.2819411
4 04 ALPES-DE-HAUTE-PROVENCE 36.30369 0.3244723 37.09966 0.4312381
5 05 HAUTES-ALPES 31.05594 -0.3794221 33.09124 -0.1064220
6 06 ALPES-MARITIMES 37.73249 0.5161209 39.76664 0.7889666
Ii E.Ii Var.Ii Z.Ii Pr.z....E.Ii..
1 -0.11976472 -0.001368720 0.02025013 -0.8319995 0.40540921
2 2.02009056 -0.054893417 0.76861540 2.3667926 0.01794298
3 0.10745221 -0.001545205 0.02285718 0.7209496 0.47094050
4 0.14142937 -0.001144240 0.02055288 0.9944955 0.31998170
5 0.04081302 -0.001564614 0.03551380 0.2248734 0.82207774
6 0.41158068 -0.002895109 0.13420143 1.1314115 0.25788193Carte des LISA
par(mfrow=c(1,2))
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Variable standardisée",
frame=T,
credits="",
arrow=F)
mypal<-brewer.pal(5,"BrBG")
mybreaks<-c(-10,-2,-1,1,2,10)
mf_map(tabres, var="Z.Ii", type="choro",
breaks=mybreaks,
pal = mypal,
leg_title = "Z-score",
leg_val_rnd = 1
)
mf_layout("LISA",
frame=T,
credits="",
arrow=F)
Typologie de Moran
# create a new variable identifying the moran plot quadrant for each observation, dismissing the non-significant ones
tabres$typ <- NA
# high-high quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-high"
# low-low quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-low"
# high-low quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-low"
# low-high quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-high"
# non-significant observations
tabres[tabres$Pr.z....E.Ii.. > 0.05, "typ"] <- "not signif."
tabres$typ <- as.factor(tabres$typ)
tabres$typcol<-NA
tabres$typcol <- case_when(tabres$typ=="high-high" ~ "red",
tabres$typ=="low-low" ~ "blue",
tabres$typ=="high-low" ~ "pink",
tabres$typ=="low-high" ~ "lightblue",
tabres$typ=="not signif." ~ "gray90"
)
mf_map(tabres, type="base", col=tabres$typcol)
mf_layout("Typologie de Moran",
frame=T,
credits ="")
mf_legend(type="typo",val = c("haut-haut","bas-bas","haut-bas","bas-haut","non sign."),pal = c("red","blue","pink","lightblue","gray90"),title = "Types")
C. ECHELLE DES CIRCONSCRIPTIONS
On passe à l’échelle des circonscriptions législatives qui sont de niveau infradépartemental. On exclue le département de Paris (75) dont les données sont imparfaites.
Préparation des données
### Fonds de carte
map<-st_read("data/elect2024/map_circ.shp", quiet=T) %>%
filter(code_dpt!=75) %>%
select(code=ID, geometry) %>%
st_transform(2154)
### Données
don<-readRDS("data/elect2024/don_circ.RDS") %>%
filter(code_dpt!=75) %>%
select(code=ID, nom=ID, 10:20)
### Jointure
mapdon<-left_join(map[,1],don)
# Table de voisinage
map_nb<-spdep::poly2nb(mapdon,row.names = mapdon$code)
# Table de poids
map_nb_w<-nb2listw(map_nb)
summary(map_nb_w)Characteristics of weights list object:
Neighbour list object:
Number of regions: 517
Number of nonzero links: 2776
Percentage nonzero weights: 1.038576
Average number of links: 5.369439
Link number distribution:
1 2 3 4 5 6 7 8 9 10 11
3 18 64 79 116 100 73 46 11 6 1
3 least connected regions:
17001 56005 59013 with 1 link
1 most connected region:
42004 with 11 links
Weights style: W
Weights constants summary:
n nn S0 S1 S2
W 517 267289 517 208.7879 2132.706# Carte de voisinage
par(mfrow=c(1,1))
coo<-st_coordinates(st_centroid(mapdon))
mf_map(mapdon, type="base",col="lightyellow")
mf_layout("Carte des liens de voisinage", frame=T)
plot.nb(map_nb,coords = coo,add = T,col = "red",points = F)
#mf_label(mapdon, var="code", col="blue",cex=0.6,halo = T, bg="white",r = 0.1)Choix d’une variable Y
sel<-mapdon %>% select(code,nom, Y=varY) %>%
mutate(Y_std = as.numeric(scale(Y)) ) par(mfrow=c(1,2))
# Carte de Y
mf_map(sel, var="Y", type="choro",
nbreaks = 6,
lwd=0.1,
leg_title = "Quantiles",
leg_val_rnd = 1)
mf_layout("Y",
frame=T,
credits="",
arrow=F)
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
lwd=0.1,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Y_std",
frame=T,
credits="",
arrow=F)
Valeurs de voisinage
sel$Y_lag<-as.numeric(lag.listw(map_nb_w,sel$Y))
sel$Y_std_lag<-as.numeric(lag.listw(map_nb_w,sel$Y_std))
#mapsel$Y_std_lag<-lag.listw(map_nb_w,mapsel$Y_std)
head(st_drop_geometry(sel)) code nom Y Y_std Y_lag Y_std_lag
1 33004 33004 30.57356 -0.1781933 28.62401 -0.3751887
2 38001 38001 15.17156 -1.7345108 24.45730 -0.7962199
3 59010 59010 34.87720 0.2566749 25.39481 -0.7014880
4 33007 33007 19.95650 -1.2510096 23.81798 -0.8608204
5 01001 01001 37.79588 0.5515966 37.05787 0.4770237
6 01002 01002 36.73677 0.4445777 33.63850 0.1315078Indice et diagramme de Moran
spdep::moran.test(sel$Y,map_nb_w, alternative = "two.sided")
Moran I test under randomisation
data: sel$Y
weights: map_nb_w
Moran I statistic standard deviate = 26.255, p-value < 2.2e-16
alternative hypothesis: two.sided
sample estimates:
Moran I statistic Expectation Variance
0.7284488328 -0.0019379845 0.0007739103 par(mfrow=c(1,1))
moran.plot(x=sel$Y_std,
listw=map_nb_w,
pch=20,
main ="Diagramme de Moran",
xlab = "Variable observée (Y)",
ylab= "Variable de voisinage (Ylag)",labels = F)
text(sel$Y_std,
sel$Y_std_lag,
sel$nom,
cex=0.5,
col="red",
pos=1)
text(-1,-0.3,"Low-Low",col="grey30")
text(-1,0.3,"Low-High",col="grey30")
text(1,-0.3,"High-Low",col="grey30")
text(1,0.3,"High-High",col="grey30") 
Calcul des LISA
locm<-localmoran(sel$Y,map_nb_w,alternative = "two.sided")
sel2<-as.data.frame(locm)
tabres<-cbind(sel,sel2)
head(st_drop_geometry(tabres)) code nom Y Y_std Y_lag Y_std_lag Ii E.Ii
1 33004 33004 30.57356 -0.1781933 28.62401 -0.3751887 0.06698569 -0.0000616558
2 38001 38001 15.17156 -1.7345108 24.45730 -0.7962199 1.38372844 -0.0058417792
3 59010 59010 34.87720 0.2566749 25.39481 -0.7014880 -0.18040333 -0.0001279258
4 33007 33007 19.95650 -1.2510096 23.81798 -0.8608204 1.07898156 -0.0030388720
5 01001 01001 37.79588 0.5515966 37.05787 0.4770237 0.26363456 -0.0005907915
6 01002 01002 36.73677 0.4445777 33.63850 0.1315078 0.05857873 -0.0003837836
Var.Ii Z.Ii Pr.z....E.Ii..
1 0.004500391 0.9994395 0.31758186
2 0.996965365 1.3916835 0.16401828
3 0.021957452 -1.2165935 0.22375891
4 0.310831367 1.9407654 0.05228674
5 0.060577565 1.0735409 0.28302851
6 0.028004175 0.3523422 0.72458168Carte des LISA
par(mfrow=c(1,2))
# Carte de Ystd
mypal<- brewer.pal(6, "RdYlBu")
mybreaks<-c(-10,-2,-1,0,1,2,10)
mf_map(sel, var="Y_std", type="choro",
breaks=mybreaks,
pal = mypal,
lwd=0.1,
leg_title = "moy. & ec.type",
leg_val_rnd = 1
)
mf_layout("Variable standardisée",
frame=T,
credits="",
arrow=F)
mypal<-brewer.pal(5,"BrBG")
mybreaks<-c(-10,-2,-1,1,2,10)
mf_map(tabres, var="Z.Ii", type="choro",
breaks=mybreaks,
pal = mypal,
lwd=0.1,
leg_title = "Z-score",
leg_val_rnd = 1
)
mf_layout("LISA",
frame=T,
credits="",
arrow=F)
Typologie de Moran
# create a new variable identifying the moran plot quadrant for each observation, dismissing the non-significant ones
tabres$typ <- NA
# high-high quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-high"
# low-low quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-low"
# high-low quadrant
tabres[(tabres$Y_std >= 0 &
tabres$Y_std_lag <= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "high-low"
# low-high quadrant
tabres[(tabres$Y_std <= 0 &
tabres$Y_std_lag >= 0) &
tabres$Pr.z....E.Ii..<0.05, "typ"] <- "low-high"
# non-significant observations
tabres[tabres$Pr.z....E.Ii.. > 0.05, "typ"] <- "not signif."
tabres$typ <- as.factor(tabres$typ)
tabres$typcol<-NA
tabres$typcol <- case_when(tabres$typ=="high-high" ~ "red",
tabres$typ=="low-low" ~ "blue",
tabres$typ=="high-low" ~ "pink",
tabres$typ=="low-high" ~ "lightblue",
tabres$typ=="not signif." ~ "gray90"
)
mf_map(tabres, type="base", col=tabres$typcol)
mf_layout("Typologie de Moran",
frame=T,
credits ="")
mf_legend(type="typo",val = c("haut-haut","bas-bas","haut-bas","bas-haut","non sign."),pal = c("red","blue","pink","lightblue","gray90"),title = "Types")